A new classifier based on information theoretic learning with unlabeled data
β Scribed by Kyu-Hwa Jeong; Jian-Wu Xu; Deniz Erdogmus; Jose C. Principe
- Publisher
- Elsevier Science
- Year
- 2005
- Tongue
- English
- Weight
- 348 KB
- Volume
- 18
- Category
- Article
- ISSN
- 0893-6080
No coin nor oath required. For personal study only.
β¦ Synopsis
Supervised learning is conventionally performed with pairwise input-output labeled data. After the training procedure, the adaptive system's weights are fixed while the testing procedure with unlabeled data is performed. Recently, in an attempt to improve classification performance unlabeled data has been exploited in the machine learning community. In this paper, we present an information theoretic learning (ITL) approach based on density divergence minimization to obtain an extended training algorithm using unlabeled data during the testing. The method uses a boosting-like algorithm with an ITL based cost function. Preliminary simulations suggest that the method has the potential to improve the performance of classifiers in the application phase.
π SIMILAR VOLUMES
We recently described a new family of oxide anion conductors with a structure based on [Bi 12 O 14 ] columns (J. Solid. State Chem. 122, 394 (1996)). The parent compound of this series can be formulated as Bi 26 Mo 10 O 69 and formation of a solid solution, in the Bi 2 O 3 -MoO 3 binary system, in t